Shangyin Tan

My 'Hair Loss' when getting a PHD is 19,564, which is much less than actual. I suppose the hair loss would be between 100k to 200k. (Assume PHD is completed...

Add missing latent effect for top-lambda node

Easier api for transforming modules with assertions

6

Now, it is possible to write `assert_transform_module(module)` to transform a module to handle suggestions and assertions. The default behavior is implemented by the `backtrack_handler`, which automatically backtracks for failing assertions...

Multi-line input incorrectly handled as single line

1

```python import dspy turbo = dspy.OpenAI(model="gpt-3.5-turbo", max_tokens=500) dspy.settings.configure(lm=turbo, trace=[], temperature=0.7) class GenerateQuestion(dspy.Signature): question = dspy.InputField() answer = dspy.OutputField() prog = dspy.Predict(GenerateQuestion) question = """ This is a dummy multi-line question....

Fix assertion bug with new_signature

2

Co-authored-by: chiragshah285 @chiragshah285

TODO - [x] Test - [x] DSPy style cache - TODO: handle non 400 return code

In some rare cases, some of the fields of a DSPy signature might have name conflicts with reserved attributes we use internally for signatures. For example, the following signature will...

If we define `ChainOfThought("question -> answer")`, the default `signature.instructions` does not get updated when the CoT module does `signature.prepend("reasoning", …)`, resulting in the following instruction in the system message “Given...

This PR adds support for terminus for token counts for failed/partial trajectories. We can get rid of `token_counts` in `AgentResults`, too? Known issue: 1. Terminus-2's summarization call is not counted....